On Thu, Aug 23, 2012 at 10:39 PM, Stephen E. Baker <baker.stephen.e@gmail.com> wrote:
On 23/08/2012 4:14 PM, Felipe Contreras wrote: [snip]
Is systemd ready? Where is the evidence?
https://www.archlinux.de/?page=PackageStatistics shows that about 14% of arch users who are using pkgstat have systemd installed. It is not default and not depended on by anything, so that means a sizable portion of the community has tried it. The bug tracker isn't being flooded with critical bugs against it so it must work for the majority of archers who are using it.
Not exactly. People might have installed it and never tried it (not very likely); or tried it, found problems, and disabled it (not so unlikely) (I did that). So it's not correct to say that it "must" work for the majority of users who are using it, but rather that it *probably* works for the majority of users who have tried it. That being said I see many problems in arch-general that are not reported in bugzilla, so I wouldn't trust bugzilla statistics blindly. And finally, the 14% that have tried it are probably early adopters, and they are not the same as the rest of the 86% of the user-base. Maybe they had problems, but the managed to solve them themselves. Maybe the rest of the 86% won't be so lucky.
Is that evidence? If not, what would constitute evidence?
That's evidence all right, but I would consider it weak evidence. Stronger evidence would be that RHEL has adopted it, or that the bug count in Fedora and openSUSE are low (they aren't), or that at least 25% of Arch Linux users have tried it. Any of those would persuade me. But that was only one of my questions, the other one would be: What are the benefits vs. advantages of systemd? It would be nice of such analysis was made publicly, but if the previous question is answered (systemd is ready), this question is not as important -- the most important thing is not to break user experience (or have a very good reason to risk doing that). Cheers. -- Felipe Contreras